When is a pesthouse not a pesthouse? When it’s been run through the Library of Congress subject heading ringer – then it’s a hospital.

The Counting the Dead project is currently developing a set of XML/TEI encoded texts for use in a future full-scale archive. As a fledgling project, however, that full-scale implementation is a future goal. In the short term, we’re creating an Omeka-based smaller archive and developing a couple of proof-of-concept style exhibits and collections. We want to begin to demonstrate the rich interplay between numerate, poetic, artifactual, etc modes of plague commemoration.

In working with Omeka, we are using the basic set of Dublin Core fields as a metadata stream (this is somewhat parallel to what we do in our TEI Headers, but not identical).

A quick grab of a Dublin Core data entry screen in Omeka

To supplement basic training on metadata, I worked with Allegra Gonzales Swift, our Digital Initiatives Librarian to develop what she calls a “data dictionary.” I was entirely new to this genre, but it functions as a kind of key for a database. Wikipedia puts it rather well: “This typically includes the names and descriptions of various tables and fields in each database, plus additional details, like the type and length of each data element. There is no universal standard as to the level of detail in such a document, but it is primarily a weak kind of data.” [1]

As we began to refine our dictionary, it became clear to me that the “subject” field was a potential quagmire. While we will be using Library of Congress subject headings, that list is extraordinarily long and difficult to quickly navigate. While I’m exploring the efficacy of an auto-complete plugin for Omeka, I want to provide a smaller controlled vocabulary for this particular field. To that end, we are developing a sub-dictionary for LOC subject terms.

Allegra asked that I generate a list of subject terms for our texts, which she and a student would then transform into official LOC terms. It seemed like a simple enough operation. Among my chosen subject terms were “pesthouse,” “receipts or recipes,” and “royal proclamations.”  The LOC translations: “Communicable diseases – hospitals,” “plague vaccines,” and “delegated legislation” or “executive orders.” I suspect that early modernists reading this will spot obvious issues. The hospital, as such, was not an active entity in early modern England. While there are active societies of physicians and it was possible to receive care through some religious organizations, it wasn’t until the 18th century that what we know as the ‘hospital’ came into being. Perhaps more importantly, the pest house was a building designated by decree to forcibly detain/contain those who were infected. It wasn’t a place for treatment; it was a place in which to die.

A similar sort of anachronism is at work with using “plague vaccine” for “recipe”- an Englishwoman or man would have loved, I’m sure, to have had a vaccine.  What they did instead, in one particularly peculiar instance, was to pluck feathers from the butt of a live chicken and apply that to oozing sores. A vaccine it was not. While the practice of inoculation (smallpox) is evident in India and China in the 17th century, such medicinal practices did not make their way in time to help with the London outbreak of 1666. The first plague vaccine was a late 19th century creation.

The final set — “delegated legislation” or “executive orders” for “royal proclamation” — doesn’t smack of quite the same anachronism, but it feels wrong. Sure, a proclamation is a piece of delegated law, but the specificity of ‘royal’ seems important to me. Likewise for the difference between “legislation” and “proclamation.”

These are just a few examples and there are plenty of cases where the terminology is perfectly appropriate. In some cases, like that of “hospital,” we’re moving forward using the LOC subject term because it includes pesthouses in the documentation. But it still doesn’t feel right. In other cases, we’ll be using capacious subject terms like “Plague–Treatment–Early works to 1800,” which helps us avoid situations that feel too anachronistic but it sacrifices a lot in terms of nuance.

Perhaps this is how it must or should be. As I mentioned at the outset, we began this process in order to narrow the already large vocabulary of the subject headings. Having historical sub-sets isn’t going to address the size problem. At the same time, there are clear benefits to seeing the pesthouses, for example, as part of the history of hospitals and corporate care.

I’m pretty new to the official languages of archivists and librarians and I clearly carry the researcher’s desire for a special kind of historicized precision into this exercise. We’re also just a week into this game and perhaps we haven’t looked hard enough. But I can’t help thinking that there’s a lot of work being done to our sense of the subject of our texts as they are subjected to LOC authority.

Feel free to take a look at the Counting the Dead Data Dictionary, which is an in-progress document right now.

In response:

Posted at the request of Allegra, our digital initiatives librarian as a response:

I think what we do is look for some standardized terms but don’t try to shoe-horn it in if it doesn’t work.

I was thinking of this this morning or last night, all a blur. Jacque would know what would make sense in her field. I think we can make a mash-up, as it were, of the LCSH if they are right for the project and terms that make sense. Catalogs care but increasingly, harvesters and aggregators of data are mashing up a great mix of terms form a variety of sources. So pristine LCSH gets muddy.

And LC can be behind the times and even politically incorrect. Case in point, Tony Crowley’s issue with describing his collection of Northern Ireland Murals documenting the Troubles is that the LC terms are Brit-centric, crown-o-filic. There is a LCNAF for the town of Derry but it is “Londonderry.” If Tony wanted to show his face in Ireland again, he needed more agnostic terms. Description should not be dangerous.

But, if you look at his subject terms, most make sense but others – and especially if you put them in a blender with unrelated data, unrelated collection content or a universal catalog – really are problematic. “1690” “ANL” “July 12” Think globally, act locally and assign subject terms accordingly.